In a recent study in JAMA Internal Medicine, Tyler VanderWeele, professor of epidemiology at Harvard T.H. Chan School of Public Health, dived into a rarely studied data question within the long-running Nurses’ Health Study and made a startling discovery: Women who attended religious services more than once a week had a 33 percent lower risk of dying during the study period, and a lower risk of dying from cardiovascular disease and cancer, compared with those who never attended religious services. The study generated scores of media stories and a flood of interview requests.
VanderWeele earned bachelor’s degrees in mathematics and in philosophy and theology from the University of Oxford. He grew up in a Protestant household, was later drawn to the Anglican Church, and is now a practicing Roman Catholic. He talked about his research with Madeline Drexler, editor of Harvard Public Health.
Q: In this paper, why did going to church so strongly lower women’s risk of dying during the study period?
A: I think it is because it affects so many different aspects of a person’s life. One is social support, both emotional and material, which is important for health. Another is that the social and behavioral norms at religious services strongly discourage smoking. Another mechanism is that religious service attendance helps cultivate an optimistic outlook on life and appears to lower depression.
There probably are other important mechanisms as well. One potential pathway, which we’re looking at now in a different data set, is that services give a sense of meaning and purpose in life; at the heart of many religious systems is an understanding of what life is about, what one is seeking, how to make sense of what is happening in life, and what are the ends for which one should be striving. Another plausible mechanism that has been suggested is self-discipline—by attending religious services and following the practices and prescriptions of the group you’re participating in, you develop self-discipline more generally, and this is good for many outcomes in life, including healthy.
Q: This research came out of a kind of scientific serendipity. Can you describe that?
A: As an academic, one tries to look for big questions and open questions—questions where you can really contribute. And being able to contribute requires good data. The Nurses’ Health Study has been used for countless high-profile studies. But no one seems to have ever made use of the service attendance data, which was originally collected as part of a measure of social support. I happened to learn about this data in 2011, when I got my first grant from the Templeton Foundation to look at religion and public health.
After a department meeting, Eric Rimm [professor of epidemiology and nutrition] came up to me and said, “You know, Tyler, I think somewhere buried in the Nurses’ Health Study is a question about service attendance. No one’s ever used it.” I looked into it, and not only was the question there, but it was there every four years since 1992. That meant we could do nuanced work on the direction of causality: Is it that only those who are healthy are able to attend religious services, or is it that religious services enable those who attend to survive? We were able to control for the former to provide evidence for the latter.
Q: You describe religious service attendance as a social determinant of health. That term is usually applied to such variables as gender or race or socioeconomic status. Does religious service attendance really belong in the same category?
A: In public health, exposures which are both common and have large effects are considered of importance. Religious service attendance is common—about 40 percent of Americans self-report attend- ing services once a week or more. And our research has shown that the effects are often quite large as well. So from that perspective, it is a powerful social determinant of health.
For many Americans, religion and spirituality constitute important markers of social identity. We would never think to omit race or socioeconomic status or gender from our analyses; even if those aren’t the focus of our research, we always control for them as possible confounders. But we almost never control for religious service attendance—even though its effects appear to be quite pronounced. I think it is important that religious participation be considered more often within public health, and I now teach a course on this topic at the School.
Q: You have suggested that church service attendance is not simply the equivalent of a bingo game or the bowling league in promoting health, but that its effects stem from something beyond the purely social component. What exactly is that something else?
A: The purely social aspect is not irrelevant. But I suspect that with a bingo game, the effects are much smaller than is the case with social participation where you also have a shared set of values and beliefs, a shared powerful religious experience of the divine or the transcendent, a shared set of behavioral norms and mutual reinforcements to follow these, and a common vision or purpose in life.
Q: Science and religion are often at cross-purposes. Have you gotten push-back from your colleagues on this work?
A: Thus far, no. People seem to be interested in the research itself. This sort of work doesn’t in any way prove that these religious systems are correct. What it does suggest is that they do contribute valuably to health.
Q: The word “health” comes from an Old English word that means wholeness, being whole. It’s also related to the root word for “holy.” Do you think these linguistic roots are meaningful?
A: To me, they suggest that we would benefit from a broadened understanding of health—health as physical, mental, social, spiritual, and maybe societal well-being, well-functioning, flourishing. Not every study needs to pursue this broad, expansive view of what health is, but we should perhaps be studying and reflecting more on the nature of health, of wholeness, itself.
